Jeff Bezos has a great saying "Your margin is my opportunity." Essentially, it means he can out-compete almost any business in industries he enters using Amazon by undercutting the competition and doing it cheaply at massive scale.
While most hardware has migrated to the cloud, resulting in developers paying services like AWS for storage and computation as opposed to buying their own hardware, the costs for these have not decreased down to hardware margins. Instead, as hardware has migrated to cloud centers margins have become more like software margins. So whereas profit for hardware companies is typically pretty small, with services like AWS, it is closer to software level margins. Put simply, Amazon's most profitable part of the business has fat margins, which is definitely someone else's opportunity.
If one looks at storage costs on Amazon for their mid tier S3 storage, it is around 2.5 cents per gigabyte per month. When deciding whether it would be worth it for myself to mine some Filecoin, I wanted to see how much it would cost to purchase a sizable number of terabytes of hard drive space. Filecoin is a cryptocurrency that is used to pay for storage space on IPFS, a global filesystem that allows storage of large amounts of data. And mining it entails letting other people store data on your hard drive. [Ed note: Investing in cryptocurrencies or tokens is highly speculative, and the market is largely unregulated. Anyone considering it should be prepared to lose their entire investment.]
I can buy an 8TB hard drive for $169 on Newegg. That means I am paying 2.1 cents per gigabyte for that hardware. When I first started writing this article a few weeks ago, the cheapest 8TB hard drive I could find was $229, which goes to show just how rapidly storage prices are dropping per TB. The hardware itself, of course, has value, but for simplicity let's just assume it is now immediately worthless. I can undercut Amazon by about 60% and charge 1.05 cents per gigabyte of storage and get my money back within about two months. Alternatively, I can undercut them by 80% and price my storage at half a penny per gigabyte and get my money back in just under six months. Then I can double it in a year, quadruple it in two years, and so forth.
That assumes Filecoin mining is relatively inefficient and no one else figures this out. However, I imagine it will rapidly converge to efficiency levels similar to Bitcoin and Ethereum mining, which means prices will be about what it takes to pay off hardware costs, or in other words, break even, in 2 years or so. That implies that storage costs of around about one-tenth of a penny per gigabyte per month are feasible on Filecoin. This analysis also ignores the fact that storage is a relatively sunk and underutilized cost. Everyone has hard disk space, and most people do not use all their space, which means once people realize they can easily use some of that excess space to mine Filecoin, rates may drop even further.
With Bitcoin mining, however, you are solely buying ASIC chips to mine Bitcoin, and they do not do anything else. For Ethereum, mining uses your graphics card, which has an opportunity cost associated with it; in other words, you cannot use your graphics card for much else while it is mining Ethereum. With Filecoin, on the other hand, it is a unique scenario because your hard drive is usable for both mining Filecoin and regular usage at the same time.
This analysis also ignores bandwidth costs, which on Amazon cost about 9 cents per gigabyte for retrievals of data stored on S3. I imagine that will be undercut as well. In San Francisco, I can get gigabit fiber with "unlimited" data for about $40 a month, which means I could offer high performant storage on Filecoin for certain locales nearby. If Filecoin truly takes off and builds a strong network, we could see an enormous proliferation of low-cost storage, and that is entirely ignoring the benefits of content based addressing, the global CDN nature of it, and all the other benefits it has to offer.